Please, try to write good commit messages. Do your best to follow these 7 rules, borrowed from Chris Beams, plus 1 extra rule:
- Separate subject from body with a blank line
- Limit the subject line to 50 characters
- Capitalize the subject line
- Do not end the subject line with a period
- Use the imperative mood in the subject line
- Wrap the body at 72 characters
- Use the body to explain what and why vs. how
- (Extra) Prefix the subject line with the component that's modified
If you wish to know why we follow these rules, please read Chris Beams' blog entry, linked above.
Rule number 8 allows us to quickly gauge if a given commit is relevant to what we're looking for when skimming the log. It adds consistency and simplifies the message. For example
ctl-client: Print trailing newline for events
is better than
Print trailing newline for events in ctl-client
Example:
ctl-client: Print trailing newline for events
If someone wants to parse this instead of using jq, a trailing
newline delimits the end of the event.
This project follows the the Linux kernel's style guide as far as coding style is concerned, with the following exceptions:
- When declaring pointer variables, the asterisk (
*
) is placed on the left with the type rather than the variable name. Declaring multiple variables in the same line is not allowed. - Wrapped argument lists should not be aligned. Use two tabs instead. There is a lot of code that uses aligned argument lists in the project, but I have come to the conclusion that these alignments are not very nice to maintain.
In case you aren't familiar with Linux's coding style, here is a short summary and some examples of acceptable formatting:
- Use tabs for indenting.
- We do not align code (mostly), but when we do, we use spaces rather than tabs. This rule is not according to the Linux style guide.
- The preferred limit on the length of a single line is 80 columns.
- User-visible string such as log messages must not be split up.
- Use space after the following keywords:
if
,switch
,case
,for
,do
,while
. - Do not use space after others such as:
sizeof
,typeof
,alignof
or__attribute__
. - Do not use typedefs.
- It is allowed to use typedefs for function pointers. This rule is not according to the Linux style guide.
static int do_something(int number, const char* text)
{
body of function
}
// Single statement
if (condition)
do_this();
// Multiple statements
if (condition) {
do_this(2, "41");
do_that();
}
// Single statement if/else
if (condition)
do_this();
else
do_that();
// Multi-statement if/else
if (condition) {
do_this();
do_that();
} else {
otherwise();
}
switch (value) {
case 3:
printf("three!\n");
break;
case 5:
printf("five!\n");
break;
case 42:
printf("the answer to life, the universe and everything: ");
// fallthrough
default:
printf("%d\n", value);
break;
}
int a = b * c + 5;
char* some_text = "some text";
char* just_text = text + 5;
char t = *just_text;
char e = just_text[1];
wayvnc has a small but growing set of unit tests, which are run on every GitHub PR. To run them locally, do the following:
meson test -C build
There are also a handful of integration tests which also run on every PR. Read the integration tests documentation for more details, but to run them locally:
./test/integration/integration.sh
There is a helper script in util/valgrind.sh to aid in memory profiling of wayvnc and wayvncctl. This can help find and eliminate memory leaks.
We run a set of tests on every PR, in three different environments.
Each run ensures that the proposed code change:
- Builds successfully
- Passes all unit tests
- Passes all integration tests
And does so in 3 different environments:
- Ubuntu as a github action
- Arch Linux as a sourcehut build
- FreeBSD as a sourcehut build
All pull requests must contain the following sentence in the description: I have read and understood CONTRIBUTING.md.